TRAITEMENT DES TROUBLES COMPORTEMENTAUX CHEZ LES MALADES ATTEINTS DE MALADIE D’ALZHEIMER

Pr Olivier SAINT JEAN HEGP Paris

INDICATIONS A TRAITER SELON

 Caractère d’urgence  Ancienneté des troubles  Répercussion sur la vie quotidienne  Répercussion sur les aidants  Modalités de début du troubles  Eventuel facteur déclenchant  Confusion  Déclenchement multifactoriel

LES PRATIQUES PROFESSIONNELLES

  QUESTIONNAIRE POSTAL AUPRES DE TROIS GROUPES DE PROFESSIONNELS EN CAROLINE DU NORD   GERONTO-PSYCHIATRES   GENERALISTES   NEUROLOGUES

  CAS CLINIQUE DE MALADE DEMENT AGITE A SON DOMICILE

COLENDA et al, JAGS 1996; 44: 1375-79

REPONSES

COLENDA et al, JAGS 1996; 44: 1375-79

GERONTO-PSYCHIATRE

GENERALISTES NEUROLOGUES

NEUROLEPTIQUES 42,60% 40,70% 50%

BENZODIAZEPINES 9,80% 17% 3,30%

ANTIDEPRESSEURS 3,30% 8,40% 3,30%

VISITE A DOMICILE 23% 15,30% 13,30%

INTERVENTION 13,10% 10,20% 16,70%

HOSPITALISATION 3,30% 0 0

THERAPIE 3,30% 1,70% 3,30%

MOTIFS DES CHOIX THERAPEUTIQUES

  EFFICACITE RECONNUE DU MEDICAMENT   EFFETS SECONDAIRES   EXPERIENCE PERSONNELLE →  EXPLICITENT 28 % DE LA VARIANCE DE

LA DECISION

COLENDA et al, JAGS 1996; 44: 1375-79

LES PREUVES D’EFFICACITE DES MEDICAMENTS  Par des essais randomisés contre placebo

  ISRS   IACE   Quelques molécules diverses

 Par des essais ouverts   Diverses molécules   Quelques molécules folkloriques

 Par le consensus d’experts et la pratique professionnelle   Méprobamate

ESSAIS RANDOMISES

IRS IACE Thymorégulateurs

L’ambiguïté de la littérature

  Le problème de la mesure du résultat   L’hétérogénéité des troubles mal rendue par

les échelles globales   Les qualités métrologiques des échelles   La durée des études

  RCT spécifiques de faible durée   Versus les essais des IACE

IRSS

Auteurs Diagnostic (n=)

Critères d’inclusion de

SCPD

Traitements Durée Efficacité

Petracca et al, 2001

Alzheimer n=41

Dépression majeure ou

mineure

Fluoxetine vs placebo

6 semaines

Pas de supériorité sur le placebo évaluation

avec l’échelle d’Hamilton et la CGI

Taragano et al

(1997)

Alzheimer n=37

Dépression majeure

Fluoxetine vs amitriptiline

6 semaines

Amélioration similaire sur

l’échelle d’Hamilton Katona et al (1998)

Démence légère ou modérée

n=198

Dépression majeure ou

mineure

Paroxetine vs imipramine

8 semaines

Amélioration similaire

Inefficacité des IRS dans la dépression du déclin cognitif

Articles

www.thelancet.com Vol 378 July 30, 2011 407

given mirtazapine compared with sertraline, but these diff erences did not persist to 39 weeks (table 4).

Our fi ndings did not diff er in subgroup analyses examining outcomes by baseline depression severity (CSDD score 8–11 vs ≥12). All but eight participants (one in the placebo group, three in the sertraline group, and four in the mirtazapine group) met criteria for categorical diagnosis of depression in Alzheimer’s disease as per Olin criteria.24 Sensitivity analyses with the Olin criteria as a moderator were not appropriate because of the low frequency of participants who did not meet Olin criteria. However, this gives reassurance of the clinical signifi cance of the depression in dementia investigated here.

Carers whose relatives were receiving placebo had higher quality of life scores at 13 weeks (SF-12 mental component score) and higher mental health scores (GHQ-12) than did those on sertraline (table 4). Finally, carers of participants in the mirtazapine group had higher quality of life scores (SF-12 mental component score) at 13 weeks than did the carers of participants in the sertraline group. However, these diff erences did not persist at 39 weeks.

Table 5 shows adverse reactions by week 39. 29 of 111 participants (26%) in the placebo group had adverse reactions, compared with 46 of 107 (43%) in the sertraline group (p=0·010) and 44 of 108 (41%) in the mirtazapine group (p=0·031; overall p value for placebo vs either drug 0·017). Gastrointestinal reactions were most common with sertraline (usually nausea) and psychological reactions were most common with mirtazapine (usually drowsiness and sedation). At 13 weeks, there were 15 serious adverse events in the placebo group of which three (20%) were rated severe, compared with 12 in the sertraline group (eight severe [67%]) and 14 (10 severe [71%]) in the mirtazapine group. Overall, the number of serious adverse events reported did not diff er between groups but more of these events were severe in those on antidepressants compared with placebo (p=0·003). Mortality did not diff er between groups (fi ve deaths in each group by 39 weeks).

DiscussionOur trial has negative fi ndings but important clinical implications. Analysis of the data suggests clearly that antidepressants, given with normal care, are not clinically eff ective when compared with placebo for the treatment of clinically signifi cant depression in dementia. This fi nding implies a need to change the present clinical practice of prescription of antidepressants as the fi rst-line treatment of depression in dementia caused by Alzheimer’s disease (panel).

Our study had limitations. First, drop outs might introduce bias if those lost to follow-up had a diff erent response to the interventions or placebo compared with those completing the trial. However this was a pragmatic trial with few exclusions designed to emulate clinical populations, and rates of disengagement were much the

same as in clinical settings. Strenuous eff orts were made to follow up and obtain outcome data for all participants who were randomly allocated to treatment and who defaulted from either the trial drug or services.

Second, we had to revise the target sample size during the trial. However, the new target was set with the same parameters as the prestudy calculations and we recruited 326 (96%) of a target of 339. Notwithstanding, our study is the largest ever randomised trial of depression in dementia with unequivocal fi ndings showing no eff ect of mirtazapine or sertraline compared with placebo. Had the pattern of change noted in recruited participants been continued, the extra precision in estimates from another 13 participants (or even achievement of the original 507) would not have generated a statistically signifi cant positive result for either antidepressant.

0 13 390

6

8

10

12

14

CSDD

scor

e

Visit (weeks)

PlaceboSertralineMirtazapine

Figure 2: Unadjusted mean CSDD scores by treatment groupLowest score is best. Error bars show 95% CIs. CSDD=Cornell scale for depression in dementia.

Placebo Sertraline Mirtazapine

n CSDD score n CSDD score n CSDD score

Baseline, mean (SD) 111 13·6 (5·2) 107 12·8 (3·6) 108 12·5 (3·7)

Week 13, mean (SD) 95 7·7 (4·1) 78 8·6 (4·9) 85 7·6 (5·0)

Week 39, mean (SD) 82 8·5 (5·5) 68 8·5 (5·5) 76 7·7 (6·2)

Mean diff erence from placebo (SE, 95% CI; p value)

13 weeks ·· ·· 173 1·17 (0·72, –0·23 to 2·58; 0·10)

180 0·01 (0·70, –1·37 to 1·38; 0·99)

39 weeks ·· ·· 150 0·37 (0·76, –1·12 to 1·87; 0·62)

158 –0·66 (0·74, –2·12 to 0·79; 0·37)

Mean diff erence from mirtazapine (SE, 95% CI; p value)

13 weeks ·· ·· 163 1·16 (0·72, –0·25 to 2·57; 0·11)

·· ··

39 weeks ·· ·· 144 1·04 (0·76; –0·45 to 2·53; 0·17)

·· ··

Comparisons of the diff erences were made at 13 and 39 weeks from the fi nal adjusted linear mixed model. CSDD=Cornell scale for depression in dementia.

Table 3: Primary outcomes of research worker rated CSDD score

Lancet 2011

IACE COCHRANE

NPI dans l’étude AD 2000

THYMOREGULATEURS

Auteurs Type de démence

(n=)

Critères d’inclusion de

SCPD

Traitements Durée Efficacité

Tariot et al

(1998)

Alz, vasc ou mixte

n=51

Agitation évaluée sur

la BPRS

Carbamazepine vs

placebo

6 semaines Sur le score total à la

BPRS

Porsteinsson et

al (2001)

Alz, vasc ou mixte

n=56

Agitation évaluée sur

la BPRS

Divalproex

sodium vs placebo

6 semaines Sur le facteur agitation

de la BPRS

NEUROLEPTIQUES

 Neuroleptiques « classiques »   Haloperidol

 Neuroleptiques atypiques   Risperidone   Olanzapine

META-ANALYSE DES ESSAIS DES NEUROLEPTIQUES

  ETAT DE LA LITTERATURE A LA FIN DES ANNEES 80   33 ESSAIS CONTRE PLACEBO (17) OU CONTRE TRAITEMENTS

DE REFERENCE (16)   EFFETS SUPERIEURS DES NEUROLEPTIQUES CONTRE

PLACEBO

SCHNEIDER LS et al, JAGS 1990 ; 48 : 553-63

META-ANALYSE DES ESSAIS DES NEUROLEPTIQUES

  EFFET « MODESTE »   AUCUN NEURLEPTIQUE N’EST SUPERIEUR A UN AUTRE   AUCUNE AUGMENTATION D’EFFET EN ASSOCIATION   EFFET THERAPEUTIQUE EQUIVALENT A L’EFFET

IATROGENIQUE

SCHNEIDER LS et al, JAGS 1990; 48: 553-63

NEUROLEPTIQUES ATYPIQUES

Auteurs Diagnostic

(n=)

Critères d’inclusion de

SCPD

Traitements Durée Efficacité

De Deyn

et al

(1999)

Alz, vasc ou

mixte n=344

Score à la BehaveAD > 8 Risperidone vs

haloperidol vs

placebo

12

semaines

Amélioration du score total

surtout sur agressivité

Katz et al

(1999)

Alz, vasc ou

mixte

n=625

Score à la BehaveAD > 8 Risperidone vs

placebo

12

semaines

Amélioration du score total

surtout sur délire

Street et al

(2000)

Alzheimer

n=206

Score aux items de la NPI,

agitation, agressivité, délire

ou hallucinations > 3

Olanzapine vs

placebo

6

semaines

Amélioration du score total

surtout sur manifestations

psychotiques

DELAI D’ACTION

RELATION EFET / DOSE

NL ATYPIQUES ET DEMENCE

  Fréquence rapportée des AVC (olanzapine versus placebo) dans des essais randomisés chez de sujets âgés atteints de déclin cognitif (Alzheimer, mixte et vasculaire)   Placebo : 2/478 (0.4%)   Olanzapine : 15/1178 (1.3 %)

  Deux fois plus de décès dans le groupe traité que dans le groupe placebo (facteurs de risques : dénutrition, pneumopathie, usage concomitant de BDZ)

  D’où la recommandation de l’AFSSAPS étendue à la risperidone et le rappel des IACE

ESSAIS AVEC LES BENZODIAZEPINES

  ALPRAZOLAM versus HALOPERIDOL A FAIBLES DOSES   DOUBLE AVEUGLE AVEC CROSS-OVER   48 MALADES DURANT 12 SEMAINES   AGE MOYEN 83 ANS   PAS DE DIFFERENCE D’EFFICACITE OU DE TOLERANCE

Christensen db et coll JAGS 1998 46 620-25

DIFFICULTES METHODOLOGIQUES

Teri et al Neurology 2000 55 1271

LES ESSAIS OUVERTS

Bêta bloqueurs Anticomitiaux THC

MOLECULES DIVERSES

 Bêta bloqueurs  Anticomitiaux (gabapentine)  THC !  Etc …

 Toujours au cours d’essais cliniques brefs rarement publiés

LES THERAPIES COMPORTEMENTALES

Luxthérapie Massage Hydrothérapie

UNE APPROCHE COMPLEXE

 Diversement applicables aux divers stades de démence

 Absence d’études sérieuses de type RCT  Une grande diversité de propositions

  Lux thérapie   Relaxation, sophrologie   Balnéothérapie   Massage chinois   Tai Chi

THERAPEUTIQUE COMPORTEMENTALE VS NEUROLEPTIQUES   COMPARAISON

  DIVERS NEUROLEPTIQUES, PLACEBO   VERSUS APPRENTISSAGE DE PRISE EN

CHARGE COMPORTEMENTALE CHEZ LES AIDANTS

  CHEZ 149 MALADES ALZHEIMER   VARIABLE DE RESULTATS

  CIBIC

Teri et al Neurology 2000 55 1271

EFFICACITE SELON CIBIC

BMT Haloperidol Trazodone Placebo Ensemble

Amélioration 13 (32%) 11 (32%) 15 (41%) 11 (31%) 50 (34%)

Stabilité 8 (20%) 7 (21%) 5 (14%) 10 (28%) 30 (20%)

Aggravation 20 (49%) 16 (47%) 17 (46%) 15 (42%) 68 (46%)

Teri et al Neurology 2000 55 1271

EFFETS SECONDAIRES

BMT N = 41

Haloperidol N = 34

Trazodone N = 37

Placebo N = 36

Parkinsonnisme 0 22 12 7

Sécheresse buccale 4 26 16 13

Tremblement 11 26 12 13

Rigidité 11 33 9 13

Asthénie 17 26 12 17

Teri et al Neurology 2000 55 1271

Recommandations ANAES

Recommandations ANAES

Recommandations ANAES